[bug] MatterThermostat: invalid initial heating setpoint (0.2°C) and setpoint clamped against absolute limits only

[lboue]

Hardware

N/A - static code review finding, reproducible by inspection on any Matter-capable board (e.g. Arduino Nano Matter)

Core version

main / 4.0.0

Arduino IDE version

N/A

Operating system

N/A

Radio stack variant

Matter

OpenThread Border Router device (if using Matter)

N/A

Issue description

DeviceThermostat is constructed with raw local temperature/heating setpoint
values of 20 (MatterThermostat.cpp:93), which is 0.2°C - below both
abs_min_heating_setpoint (700 = 7°C) and min_heating_setpoint
(1600 = 16°C) initialized right below in the same constructor
(DeviceThermostat.cpp:36-39). This mirrors a bug just fixed for
MatterWaterHeater in PR #181 - the intended value was very likely
2000 (20°C), not 20 (0.2°C).

Separately, DeviceThermostat::SetHeatingSetpointValue()
(DeviceThermostat.cpp:66) only clamps against abs_min_heating_setpoint/
abs_max_heating_setpoint, ignoring the writable
min_heating_setpoint/max_heating_setpoint limits, and computes the
changed flag before clamping, which can fire a spurious attribute
report when an out-of-range write clamps back to the current value.

Suggested fix (same pattern applied to MatterWaterHeater in PR #181):

• Use a valid initial heating setpoint (e.g. within [min_heating_setpoint, max_heating_setpoint])
• Clamp against min_heating_setpoint/max_heating_setpoint instead of the absolute limits
• Compute changed after clamping

Serial output

No response

RTT output (if using Matter)

No response

Minimal reproducer code

No response